Phee ɸ Phoh PhuZ: A Tale of Gi­ant Phage with a Furtive Tubu­lin

by Daniel P. Haeusser

A few years ago I at­tended an ASM Branch meet­ing where an in­ves­ti­ga­tor gave a talk about a metage­nomic sur­vey of oceanic bac­te­rio­phages. In typ­i­cal fash­ion for this type of study, one slide listed dozens genes of note iden­ti­fied as be­ing en­coded in phage genomes. With sur­prise I no­ticed that one of these was ftsZ, which en­codes the prokary­otic tubu­lin ho­molog re­spon­si­ble for cell di­vi­sion in most bac­te­ria and sev­eral phyla of ar­chaea. I won­dered: Why on Earth would a phage con­tain a cy­toskele­tal pro­tein? The ftsZ gene is even found in chloro­plasts and the mi­to­chon­dria of cer­tain pro­tists, but at least these or­ganelles have the evo­lu­tion­ary his­tory of hav­ing been in­de­pen­dent mem­brane-bound cells. But with a virus there is no 'cyto' in which to place any 'skele­ton'. An ob­vi­ous hy­poth­e­sis is that the phage might make use of a tubu­lin-like pro­tein within the host for it ne­far­i­ous phage re­pro­duc­tion cy­cle – but in what way?

En­ter two re­cent stud­ies (Krae­mer, et al. and Oliva, et al.) re­port­ing phage-en­coded tubu­lin ho­mologs and the start of their char­ac­ter­i­za­tion. A gen­eral model emerges for these pro­teins in di­rect­ing phage DNA repli­ca­tion within host cells and max­i­miz­ing phage pro­lif­er­a­tion. In­trigu­ingly, the phages en­cod­ing these tubu­lin ho­mologs have par­tic­u­larly large genomes, sug­gest­ing a pos­si­ble cor­re­la­tion be­tween gi­ant genome size and the need for a phage-en­coded cy­toskele­ton.

Fig. 2: The tubu­lin su­per­fam­ily with newly-iden­ti­fied PhuZ (green) clus­ter and Clostrid­ium-de­rived clus­ter (dark blue) that in­cludes the TubZ-like pro­tein of phage c‑st. Im­age from Krae­mer, et al. Not shown is the BtubAB (bac­te­r­ial tubu­lin) clus­ter.

The newly pub­lished phage tubu­lin ho­mologs fall into two phy­lo­ge­netic clus­ters (Fig. 2), each unique in the tubu­lin su­per-fam­ily. One clus­ter con­tains rep­re­sen­ta­tives in dif­fer­ent chro­mo­some, plas­mid, or phage of the genus Clostrid­ium. The tubu­lin ho­molog in the study by Oliva et al. comes from C. bot­u­linum phage c‑st, a phage al­ready known for its pro­duc­tion of the bot­u­lism neu­ro­toxin. The crys­tal struc­ture of this tubu­lin ho­molog (Fig. 3) re­veals strong sim­i­lar­ity to the TubZ fam­ily, which is in­volved in low-copy num­ber plas­mid seg­re­ga­tion in sev­eral Bacil­lus species. In this sys­tem, the C‑terminal tail of TubZ binds to TubR, an adap­tor pro­tein that also binds tubS cen­tromeric sites on plas­mid DNA.

Like canon­i­cal TubZ, c‑st phage TubZ as­sem­bles into two-stranded he­li­cal fil­a­ments in the pres­ence of GTP. In ad­di­tion to en­cod­ing TubZ and a TubR ho­molog, C. bot­u­linum c‑st phage also con­tains a novel fac­tor called TubY that Oliva et al. char­ac­ter­ize as a ro­bust in vitro mod­u­la­tor of TubZ as­sem­bly, ca­pa­ble of dis­as­sem­bling and re­shap­ing TubZ poly­mers. Be­cause the c‑st phage repli­cates as a plas­mid it is not sur­pris­ing that a TubZRS sys­tem could be in­volved. How­ever, it still re­mains to be de­ter­mined whether c‑st phage TubZ can form fil­a­ments in vivo or whether the sys­tem has sig­nif­i­cant ben­e­fits for phage re­pro­duc­tion.

Fig. 3: Com­par­i­son of tubu­lin su­per­fam­ily mem­ber crys­tal struc­tures. (A) The alpha/beta tubu­lin het­erodimer. (B) The BtubA/B het­erodimer. © FtsZ. (D) TubZ. (E) PhuZ. (F) Phage c‑st TubZ.

The sec­ond phy­lo­ge­netic clus­ter of phage-en­coded tubu­lin ho­mologs has been named PhuZ for "Phage tubulin/FtsZ". Rep­re­sen­ta­tives of this fam­ily hail from phages of the genus Pseudomonas, the PhuZ of P. chloro­raphis phage 201ɸ2–1 be­ing the sub­ject of the study by Krae­mer et al. Like TubZ, PhuZ as­sem­bles in the pres­ence of GTP into two-stranded he­li­cal fil­a­ments that look more gen­er­ally like f‑actin than a tubu­lin. How­ever, there are some no­table dif­fer­ences be­tween PhuZ and TubZ (Fig. 3). PhuZ lacks a con­served in­ter­do­main he­lix that is im­por­tant for poly­mer­iza­tion of other tubu­lin ho­mologs. In­stead, to achieve poly­mer­iza­tion, PhuZ uses a unique acidic patch at the end of its ex­tended C‑terminus. In the model pro­posed by the au­thors, six acidic amino acids of the thir­teen ex­treme C‑terminal residues of one PhuZ monomer form a 'knuckle' that nes­tles into a ba­sic patch formed by he­lices H3‑5 on the ad­ja­cent monomer, cre­at­ing an over­lap­ping pat­tern of poly­mer­iza­tion. A dele­tion of the knuckle and strate­gic point mu­ta­tions abol­ished PhuZ as­sem­bly in vitro, sup­port­ing this struc­tural model. Thus, in­stead of us­ing its C‑terminus to in­ter­act with other fac­tors as TubZ and FtsZ do, or to en­hance lat­eral in­ter­ac­tions, PhuZ ap­pears to em­ploy its tail for the self-in­ter­ac­tions of fil­a­ment as­sem­bly.

Fig. 4: GFP-tagged PhuZ as­sem­bles into fil­a­ments (per­haps mem­brane-as­so­ci­ated) when over­ex­pressed in vivo. Source of adapted im­age.

Ad­mirably, Krae­mer et al. ad­dressed the in vivo role of PhuZ. At high ex­pres­sion lev­els, GFP-PhuZ formed dy­namic fil­a­ments across the length of P. chloro­raphis cells (Fig. 4). The au­thors then low­ered GFP-PhuZ ex­pres­sion be­low the thresh­old level where fil­a­ments were ob­served and in­fected cells with phage 201ɸ2–1. As a re­sult, flu­o­res­cent PhuZ poly­mers would only form when ad­di­tional PhuZ was syn­the­sized by the na­tive phage. Time-lapse mi­croscopy re­vealed that PhuZ fil­a­ments formed and per­sisted un­til the host ly­ses.  Stain­ing the DNA re­vealed a high con­cen­tra­tion of phage-en­cap­si­dated DNA that formed a sin­gle rosette struc­ture at mid­cell with fre­quent con­tacts to the ends of the PhuZ fil­a­ments (Fig. 5). Ex­pres­sion of a PhuZ mu­tant that forms non-dy­namic fil­a­ments dis­played mis­lo­cal­ized phage DNA at the cell poles, fre­quently scat­tered the DNA into smaller nu­cleoids, and dra­mat­i­cally re­duced phage burst size. Al­though it is not yet clear how this com­pares to burst size un­der nor­mal PhuZ ex­pres­sion lev­els, it sug­gests that PhuZ fil­a­ments help to in­crease phage yield.

Fig. 5: Fil­a­ments of GFP-tagged PhuZ form on each side of a mid­cell-lo­cal­ized phage DNA nu­cleoid dur­ing phage in­fec­tion. Op­ti­cal sec­tions at 450, 600, and 900 nm show the rosette-like struc­ture of the phage DNA nu­cleoid. Source of adapted im­ages.

So, how ex­actly does PhuZ fa­cil­i­tate phage DNA repli­ca­tion and phage pro­lif­er­a­tion? One pos­si­bil­ity is that it di­rectly in­ter­acts with phage DNA, or in­di­rectly through an adapter pro­tein like TubZ does with TubR. Or the sys­tem may be even more com­plex. Stud­ies on ɸ29 phage of Bacil­lus have iden­ti­fied a small coiled-coil pro­tein, p1, which self-as­sem­bles into fil­a­ments and sheets that also ap­pears to play a sim­i­lar role in or­ga­niz­ing phage DNA repli­ca­tion. Apart from the p1 pro­tein, ɸ29 phage ex­presses ad­di­tional pro­teins in­volved in this process that bind to phage DNA and which de­pend on the host cell's MreB cy­toskele­ton for their lo­cal­iza­tion and ac­tiv­ity. Thus, PhuZ may be a sin­gle part of a larger ap­pa­ra­tus that in­cludes both other phage pro­teins and host cell fac­tors.

The cor­re­la­tion be­tween phage genome size and these phage-en­coded cy­toskele­tal pro­teins is en­tic­ing, but the mean­ing is not yet clear. Is there some in­her­ent need to bet­ter-or­ga­nize a large genome within the host com­pared to a smaller one? If so, why carry your own cy­toskele­ton rather than sim­ply make use of the one al­ready pro­vided by the host as many eu­kary­otic viruses do? Per­haps us­ing the host's cy­toskele­ton would be too prone to host-dri­ven as­sem­bly reg­u­la­tion that could thwart the phage's goal of ly­sis. Rather than fight to con­trol as­sem­bly of the host's nat­ural cy­toskele­ton, us­ing your own pro­tein when and where you want could be eas­ier, par­tic­u­larly when your genome af­fords you the space to in­clude it. As these and other ques­tions are ad­dressed, the cy­toskele­tal fam­ily will surely only con­tinue to grow, par­tic­u­larly with the amaz­ing field of bac­te­rio­phage open for ex­plo­ration.

 

Note: A sep­a­rate and more de­tailed ver­sion of this story is pub­lished as a Dis­patch with William Mar­golin.

Ref­er­ences

Krae­mer JA, Erb ML, Wad­dling CA, Monta­bana EA, Zehr EA, Wang H, Nguyen K, Pham DS, Agard DA, Pogliano J (2012). A phage tubu­lin as­sem­bles dy­namic fil­a­ments by an atyp­i­cal mech­a­nism to cen­ter vi­ral DNA within the host cell. Cell, 149 (7), 1488−1499. PMID 22726436

Oliva MA, Mar­tin-Galiano AJ, Sak­aguchi Y, An­dreu JM (2012). Tubu­lin ho­molog TubZ in a phage-en­coded par­ti­tion sys­tem. Proc Natl Acad Sci USA, 109 (20), 7711−7716. PMID 22538818

 

Daniel is a post­doc­toral fel­low in the Mar­golin Lab in the De­part­ment of Mi­cro­bi­ol­ogy & Mol­e­c­u­lar Ge­net­ics at the Uni­ver­sity of Texas, Hous­ton Med­ical School. He also teaches as an ad­junct pro­fes­sor at the Uni­ver­sity of Hous­ton-Down­town in the De­part­ment of Nat­ural Sci­ences.